OBSERVATIONS ON THE AURORA BOREALIS.

519

appears at all. To admit that the arch is a visible part of a complete ring that sur-rounds the magnetic pole of the earth, and that at the same time it crosses at rightangles the magnetic meridian of every place of observation, would be to admit a com-plete system of confusion—in fact, an absurdity. According to ITansteen’s and Bar-low’s maps, the curve of equal variation that passes through Great Britain passes alsoa little north of the Western Isles, through Newfoundland and into Hudson’s Bay ;and, in the other direction, it passes through the North Sea, the Shetland Islands, andthence almost direct north past Spitzbergen. No circular ring that could possibly beimagined to surround the north-western magnetic pole of the earth, would answerthe other parts of the hypothesis, for all the magnetic meridians of that particularcurve of equal variation. In Hudson’s Bay the magnetic meridian would be at rightangles to the magnetic meridian of Great Britain ; and in many parts of the curvethere would be such obliquities of the magnetic meridians to each other, that but very-few of them would cross tangents to the supposed ring, at right angles, and at thepoint of contact—circumstances required by the hypothesis ; or, in other words, thereare but very few magnetic meridians in this curve of equal variation that run suffi-ciently close to the north-western magnetic pole of the earth to satisfy the conditionsof the hypothesis. In Great Britain, and throughout the northern parts of the curve,the magnetic meridians pass north of the pole ; and the magnetic meridian, for thesame curve, opposite the coast of Spitzbergen, would pass the magnetic pole north-ward upwards of 10° of latitude. Westward, over the Atlantic Ocean, the magneticmeridians of this curve would approach the north magnetic pole more closely; and,in Newfoundland, the magnetic meridians would probably pass through that pole ;but on the American Continent, towards Hudson’s Bay, the magnetic meridians wouldpass the north-western pole some three or four degrees on its south side.

I have selected this particular curve of equal variation, as it stands in ProfessorBarlow’s map, because it is that which passes through the North of England, andcorresponds with the variation at Kendal (25°) when Dr. Dalton made his observa-tions. Had the selection been made on the curve of 20°, which passes through themost western parts of Spitzbergen, through Norway, France, Algiers, and the CanaryIslands, thence across the Atlantic to Nova Scotia, Canada, and Hudson’s Bay, thedeviations of the magnetic meridians from the magnetic pole would have been muchgreater, especially in Europe, where the meridians have been more exactly-ascertained than in any other part of the curve—that is, through the whole of thecurve from the Mediterranean to Spitzbergen; in which the magnetic meridians wouldcross the meridian in which the north-western magnetic pole is situated on itsnorthern side, and the magnetic meridian of Spitzbergen would cross the meridian ofthe pole 15° or more north of it. In the western portion of the curve, however, fromAlgiers to Hudson’s Bay, the magnetic meridians would run sufficiently close upon